HOW IT CONNECTS
From AI request to a controlled result.
Governora sits between your connected AI applications and the data, tools and models they use. It checks the request, selects an allowed route and records what happened. This page explains the components behind that flow.
The architecture has a provider-compatible edge, governed execution orchestrator, decision resolver, evidence planner, Capability Registry, authorization broker, tool runtime, provider router, outcome verifier, permission-aware memory and audit ledger.
Three integration modes
Managed gateway execution is authoritative. A customer-hosted connector keeps credential custody while Governora controls the plan. Claude or ChatGPT MCP companion mode is advisory because the host controls its own tool loop.
One normalized lifecycle
Provider-native requests are normalized into the governed execution contract. States move from received through planning, retrieval, ready, model running and validation to a terminal outcome.
Five authorization layers
Identity, registry definition, request envelope, authorization reference and source-system permission remain separate. A successful call requires every layer to agree.
Permission-aware memory
Reusable claims retain provenance, freshness and an authorization policy. Permission-bound evidence is re-authorized at read time and cannot broaden access.
Provider-independent control
Routing selects active models allowed for the tenant and task. Existing OpenAI and Anthropic formats remain at the edge while the internal plan stays provider-independent.
HOW IT WORKS
Four clear steps.
Each step stays connected to the same execution. This makes authority, external work and the final outcome explainable.
- 01
Receive
Authenticate actor and tenant; validate task, risk and capabilities.
- 02
Plan
Freeze decisions and classify evidence before external calls.
- 03
Execute
Retrieve only gaps and run residual model work if necessary.
- 04
Record
Attach reported outcome evidence and write the append-only receipt.
QUESTIONS
Simple questions. Direct answers.
What Governora does, what it does not do, and where it fits.
Does Governora replace our login system?
No. Your identity system stays in charge. Governora uses the approved identity and access information it receives.
Must every request use a model?
No. A valid reusable result or an MCP knowledge lookup can return governed context without making a model call.
Is this a final technical design?
No. It shows the main flow. Your network, scale, identity, storage, and support setup still need a detailed design.
Map one governed AI execution.
We will map the actor, decisions, capabilities, evidence, external calls and proof with you.
Plan a pilot review →